Many cool-season grasses are infected by endophytic fungi that grow within their aboveground tissues. These endophytes could potentially impact the dynamics of host populations. An experimental population of Lolium perenne was established in a field plot in New Jersey, USA to investigate the impact of endophyte infection and host genotype on tiller dynamics, sexual reproduction, and biomass over three growing seasons. Both infected (E+) and uninfected (E–) plants of each of nine L. perenne genotypes were monitored for tiller and flowering spike production at 2-3 month intervals. Endophyte infection intensity, assessed as the number of hyphae per microscopic field-of-view at 400X, was recorded on three occasions.

Results/Conclusions

Endophyte infection intensity varied significantly among genotypes and years, but there was no evidence of a decline in infection intensity over a 5-yr period. E– plants remained endophyte-free throughout the experiment. Tiller dynamics varied significantly among host genotypes and was affected by endophytes because E+ plants of several genotypes began to accumulate more tillers than E– plants during the third year. E+ plants also had greater aboveground biomass by the end of the field trial, but host genotype explained a far greater proportion of the phenotypic variation in tiller dynamics, final tiller number and biomass than infection status. Plant survival, the proportion of plants that flowered, flowering date, number of spikes, and mean tiller mass were not affected by endophyte infection. However, the latter three variables showed highly significant variation among host genotypes. Although other studies have demonstrated a consistent positive growth effect of endophyte infection on several grass hosts, in the relatively benign, noncompetitive field environment in this experiment host genotype accounted for far more of the variation in long-term tiller dynamics, vegetative growth, and sexual reproduction of Lolium perenne than endophyte infection. In terms of the evolutionary ecology of endophyte-infected grasses, these results imply that natural selection could readily differentiate among host genotypes, but that the selection process would not depend greatly on whether or not individuals were infected.